Reading your Meta results without a marketing degree

Reading Meta ads results comes down to one question: what did each euro buy, and why? Ads Manager shows you dozens of columns, but an owner needs about a dozen of them, read in a fixed order. Start with results and cost per result. If that number is fine, you're done for the week. If it isn't, the other columns tell you where the problem sits: in the audience, in the ad, or on your website.

Below: those columns, the two settings that change what they mean, why Meta's lead count won't match yours, and a ten-minute Monday routine.

Which Ads Manager columns does an owner actually need?

You need the columns that follow the path from "the ad was shown" to "someone paid you". Everything else is optional. Here they are in the order a person moves through them.

Meta's own help pages explain the difference between Clicks (all) and link clicks and between link clicks and landing page views if you want the exact definitions.

What is the attribution setting, and why does it change your numbers?

The attribution setting is the rule Meta uses to decide whether a lead or a sale counts as coming from your ad. It is set per ad set, and it decides how much your "results" column can include.

Under Meta's standard attribution, a conversion can be credited to the ad if it happens within 1 or 7 days after someone clicks a link in the ad (click-through), within 1 day after they merely saw the ad (view-through), or within 1 day after they engaged with it in another way, such as watching a video for a few seconds (engage-through). Meta's help pages describe 7-day click or 1-day view as the default for new ad sets, and Meta keeps adjusting how these windows work, so don't rely on memory. Add the Attribution setting column in Ads Manager and see what your own ad sets use. Meta's About attribution models and attribution settings page is the current reference.

Two practical consequences:

• View-through results are generous. Someone who scrolled past your ad and then found you on Google the same day can still be counted. That's not wrong, but it's not the same as "the ad sent them".
• Ads Manager reports by the day of the ad impression, not the day of the conversion. If someone clicked on Monday and enquired on Friday, the result is shown on Monday. Last week's numbers can still grow a little after the week is over.

What is the learning phase, and why shouldn't you touch the ad during it?

The learning phase is the period when Meta's delivery system is still working out who to show your ad set to. Per Meta's documentation on the learning phase, an ad set usually leaves it after about 50 results in the week after its last significant edit. While it's learning, costs are less stable and usually higher.

What counts as a significant edit, per Meta: any change to targeting, creative or optimisation event, adding a new ad to the ad set, changing the bid strategy, and pausing the ad set for 7 days or longer. A large budget change can count too; nudging it from €20 to €21 won't. Each significant edit sends the ad set back to the start.

If your ad set can't reach roughly 50 results a week, Ads Manager shows Learning limited. That isn't a penalty. It means the setup is too thin: too small an audience, too small a budget, or an optimisation event that happens too rarely. The usual fixes are fewer ad sets with more budget each, a broader audience, or optimising for a more frequent event.

For a small local business spending €10–20 a day on leads, 50 results a week is often out of reach. That's fine. It just means you judge on a longer window and resist the urge to edit every other day.

Why doesn't Meta's count match your own count of enquiries?

Meta counts what its pixel and lead forms can see and credit to an ad; you count people who actually got in touch. The two will never match exactly, and neither one is lying. The usual reasons for the gap:

1. Attribution windows. Meta can count someone who saw the ad and later came through another route. Your notebook just says "found us on Google".
2. Tracking gaps. Cookie refusals, browsers that block tracking and pages without the pixel mean some real enquiries never reach Meta.
3. Channels Meta can't see. Phone calls, Messenger chats that turn into calls, walk-ins and word of mouth after someone saw your ad.
4. Lead quality. A lead form submission is a result for Meta even when the phone number is fake or the person was just curious.
5. Timing. Meta dates the result by the impression; you date the enquiry by when it arrived.

The working rule: use Meta's count to compare ads and audiences with each other, and use your own count to decide whether the money is worth spending. Keep a simple sheet: date, where the customer says they found you, and whether it became a sale. After a month you'll know how Meta's number relates to yours.

A worked example: reading a small campaign

Here is a real one. Aknaässad, our own window-cleaning business in Estonia, spent €49.92 on Meta ads in high season. That brought 22 leads, and the jobs that came from them were worth about €2,380, with sales closed on site.

What the numbers say:

• Cost per result: €49.92 ÷ 22 is about €2.27 per lead. That's the number to compare the next campaign against.
• ROAS: Meta couldn't show it, because the sale happened at the customer's door, not on a website. The owner's own records show roughly €2,380 of work for €49.92 of spend. That number only exists because someone wrote the sales down.
• Context: this was high season, when people want clean windows anyway. The same ad in January would not produce the same result, and that's not a sign the ad got worse.

The lesson isn't the size of the return. The useful number came from joining Meta's lead count to the business's own sales record.

A 10-minute weekly review routine

Once a week, same day, same order. Set the date range to the last 7 days and compare with the previous 7.

1. Minute 1–2: results and cost per result. Per campaign. Did it get better, worse or stay the same? If it's in line with what a customer is worth to you, skim the rest and stop.
2. Minute 3–4: your own count. How many real enquiries and sales came in? Put them next to Meta's results in your sheet.
3. Minute 5–6: the chain. If cost per result rose, check CPM (auction), CTR (link) (ad) and landing page views vs link clicks (page). One of them moved more than the others.
4. Minute 7–8: frequency and fatigue. Is frequency climbing while CTR drops? Then the ad needs a fresh image or angle, not more budget.
5. Minute 9–10: one decision. Write down one thing to change, or decide to change nothing. "Nothing, it's still learning" is a valid decision.

Change one thing at a time. If you change the audience, the image and the budget on the same day, you won't know which one worked. Our guide to the first 30 days of marketing has more on building this kind of steady rhythm.

Common misreads

The most expensive mistakes in Ads Manager are not settings errors. They're reading errors.

• Judging an ad by CTR alone. A funny ad can get lots of clicks from people who will never buy. A plain ad with a clear price can get fewer clicks and more customers. CTR tells you about interest, cost per result tells you about business.
• Cutting ads after two days. Two days of data during the learning phase is mostly noise. Unless something is clearly broken (no delivery, a typo, a dead link), give an ad set about a week before you judge it.
• Comparing CPMs across seasons. Attention costs more when every shop is advertising. A CPM that jumps in November compared with September is the calendar, not your ad. Compare with the same weeks last year.
• Trusting a result without checking what was optimised. If the campaign objective was traffic, "results" are clicks, and a low cost per result says nothing about leads.

If you'd rather not do this reading yourself, the Pilgud Agent syncs results from your Meta account every day, tracks leads and suggests what to do next, and the Pilgud team can run the whole thing with you. Either way the numbers stay in your own account.